Detection Based On Orthogonal Matching Pursuit Algorithm And Its Application In The Spectrum Detection

Compressive Sensing (Compressed Sensing, CS) has brought a new revolution to the signal sampling theory. The characteristic that only a small amount of observations will be able to collect the compressible signal’s information efficiently determines the Compressed Sensing being applied to many fields, such as:spectrum sensing.Compression spectrum sensing technology caused widespread when put forth, because signal can be sampled below the Nyquist sampling rate and applied to spectrum sensing technology. Currently various compression spectrum sensing method has been widely concerned when proposed because it can reduce the sampling rate, and easy to implement, then some classic algorithm and improved algorithm put forward, but such algorithms are mostly carried out under conditions of sparsity known. However, in the practical, the sparsity is unknown.In order to detect the problem of spectral holes fast and accurate, this thesis discusses a Spectrum detection technology based on DOMP Algorithm. The method combines the spectrum sensing technology and compression sensing technology, to achieve much lower rate than the Nyquist sampling rate and then realize the non-destructive sampling and signal reconstruction.Firstly, the key technology of spectrum sensing has been introduced briefly, and then elaborates the present situation of spectrum sensing techniques based on CS theory has been elaborated.Secondly, the detection algorithm based on transmitter and receiver has been described, and the characteristics comparison has been made between the transmitter and receiver, and then a simulation and analysis for the energy detection algorithm has been done based on transmitters.Again, the CS theory from the signal sparse、transform observation matrix design and reconstruction algorithm designs aspects has been elaborated.On this basis, several common greedy iterative algorithms have been introduced, and a simulation experiment has been made. In practical applications, the priori hasn’t been known, then the DOMP been discussed, and a Simulation been done.Finally, the spectrum sensing technology and the algorithm of DOMP had been combined and then single-node detection under conditions of sparsity unknown had been realized. The single-point detection may be emerge a problem of the hidden terminal. In terms of some defects in reliability and practicality, so a multi-node collaborative detection method was introduced. And a comparison has been done between multi-node detection and single-node detection, and then a simulation and performance analysis has been done. The results show that multi-point detection has a higher detection rate, as well as it can solve the hidden terminal problem, shadow effects, etc bitterly than a single point.